pith. sign in

arxiv: 2605.17812 · v1 · pith:25IEUVQWnew · submitted 2026-05-18 · 💻 cs.AI

Going Headless? On the Boundaries of Vertical AI Firms

Muhammad Zia Hydari , Farooq Muzaffar This is my paper

Pith reviewed 2026-05-20 10:55 UTC · model grok-4.3

classification 💻 cs.AI
keywords vertical AIheadless architectureaccountability boundaryvalue capturerule debtinterface boundarycospecialized assetsplatform strategy
0
0 comments X

The pith

Vertical AI firms should keep accountability boundaries fixed while allowing interface boundaries to move if they want to retain value when going headless.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Vertical AI firms have traditionally bundled workflow, domain logic, and accountability in one application, but general-purpose agents now unbundle the package and push firms toward headless models that expose expertise as callable services. The paper claims this move succeeds only when the accountability boundary stays in place, because durable value capture lives in cospecialized assets such as professional signoff, regulated workflows, evidence trails, and trusted systems of record. Moving the accountability boundary along with the interface cedes that value to orchestrators that gain envelopment power through open protocols. The argument produces a three-position taxonomy of component, integrated software platform, and dual-track positions that depends on task-accountability regime rather than sector, and it introduces the notion of rule debt as the future governance burden that accrues when rules shift into prompts and agent instructions. Four principles follow for deciding boundary placement.

Core claim

Going headless is correct for some vertical AI firms and destructive for others; the outcome turns on whether the accountability boundary moves with the interface boundary, and durable value capture concentrates in cospecialized accountability assets including professional signoff, regulated workflows, evidence trails, and trusted systems of record.

What carries the argument

The distinction between the interface boundary, which can often move, and the accountability boundary, which often must not move, as the factor that determines whether architectural choices cede value capture.

If this is right

  • Firms that decompose offerings by accountability regime rather than by interface will preserve their position in the value chain.
  • Integrated software platforms reduce customer exposure to rule debt by keeping business rules and professional standards inside governed systems.
  • Dual-track approaches allow firms to expose some services to agents while retaining core accountability functions internally.
  • Avoiding dependence on a single orchestrator prevents the loss of control over evidence trails and signoff processes.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same boundary logic could be tested in less regulated verticals where accountability norms are still forming.
  • Customers could measure rule debt directly by tracking the hours spent auditing and updating agent instructions over time.
  • Platforms that currently sit at the orchestrator layer might face pressure to add their own accountability layers if the pattern holds.

Load-bearing premise

Accountability requirements in regulated domains cannot be fully transferred to general agents and open protocols without losing the cospecialized assets that protect value.

What would settle it

A documented case of a vertical AI firm that relocated its accountability boundary into agent instructions and open services yet retained or grew its share of customer spending and margins over multiple years.

read the original abstract

Vertical AI firms in accounting, law, healthcare, procurement, and similar domains historically bundled workflow, domain logic, and accountability into a single application. General-purpose AI agents are now unbundling that package, prompting founders and investors to advocate "going headless": cede the workflow and interface to agents and expose domain expertise as callable services. This article argues that going headless is correct for some firms and destructive for others, and that the latter often cede their value capture inadvertently through architectural choices that look like interface decisions. This is a boundary question, and the answer turns on distinguishing the interface boundary, which can often move, from the accountability boundary, which often must not. Drawing on Coase's theory of the firm, Eisenmann, Parker, and Van Alstyne's platform envelopment framework, and Teece's analysis of complementary assets and appropriability, the article shows that orchestrators operating through open protocols acquire envelopment power even as technical interoperability improves, and that durable value capture concentrates in cospecialized accountability assets: professional signoff, regulated workflows, evidence trails, and trusted systems of record. The article proposes a three-position taxonomy (component, integrated software platform, dual-track) determined not by sector but by task-accountability regime, and formalizes the construct of rule debt: the future governance, maintenance, and accountability burden that accrues to customer organizations when business rules and professional standards migrate from governed systems into prompts and agent instructions. Four principles follow: decompose by accountability not interface, invert the edges while retaining the core, position rule debt as the customer cost the integrated platform prevents, and avoid single-orchestrator dependence.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript claims that vertical AI firms in domains such as accounting, law, and healthcare should not uniformly 'go headless' by ceding workflows and interfaces to general-purpose agents while exposing domain logic as services. Instead, the decision hinges on distinguishing the movable interface boundary from the accountability boundary, which often must remain fixed. Drawing on Coase's theory of the firm, Eisenmann et al.'s platform envelopment framework, and Teece's analysis of complementary assets and appropriability, the paper argues that durable value capture concentrates in cospecialized accountability assets (professional signoff, regulated workflows, evidence trails, and trusted systems of record). It proposes a three-position taxonomy (component, integrated software platform, dual-track) based on task-accountability regime rather than sector, introduces the construct of 'rule debt' as the future governance burden from migrating rules into prompts, and derives four principles: decompose by accountability not interface, invert the edges while retaining the core, position rule debt as the customer cost the integrated platform prevents, and avoid single-orchestrator dependence.

Significance. If the boundary distinction holds, the paper supplies a useful conceptual lens for strategic architecture choices in vertical AI, integrating established economic and strategy theories with AI unbundling dynamics. The formalization of rule debt as an explicit customer cost is a constructive addition that highlights governance and maintenance implications. Credit is due for grounding the analysis in well-known frameworks (Coase, platform envelopment, Teece) without introducing free parameters or ad-hoc entities beyond the defined rule debt construct, and for framing the taxonomy around task regimes rather than sectors to enable cross-domain application.

major comments (2)
  1. [section proposing the three-position taxonomy and the four principles] The central claim that the accountability boundary often must not move (while the interface boundary can) is load-bearing for the taxonomy and the four principles, yet the manuscript supplies no explicit operational criteria, decision rules, or measurement approach for classifying a given task or workflow into a regime where accountability must remain fixed versus one where it can migrate into prompts or agent instructions. This renders the distinction an assertion rather than a testable discriminator.
  2. [discussion applying Eisenmann, Parker, and Van Alstyne's platform envelopment framework] The argument that orchestrators operating through open protocols acquire envelopment power even as technical interoperability improves rests on the platform envelopment framework, but the manuscript does not address how the accountability boundary interacts with envelopment dynamics in concrete cases (e.g., whether regulated workflows in healthcare create a natural barrier or merely delay envelopment).
minor comments (2)
  1. [formalization of rule debt] The definition of rule debt could be sharpened by distinguishing it more explicitly from related concepts such as technical debt or compliance debt, with a brief reference to prior literature on governance burdens in IT systems.
  2. [abstract] The abstract previews the taxonomy and principles but does not indicate how many illustrative cases or counterexamples (if any) appear in the full text; adding one sentence on the empirical or illustrative content would improve reader expectations.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive report and for recognizing the paper's contributions in linking Coase, platform envelopment, and Teece to vertical AI architecture choices. We agree that the accountability-interface distinction and its implications for envelopment merit further operational detail. We respond to each major comment below and will incorporate the suggested elaborations in revision.

read point-by-point responses

  1. Referee: [section proposing the three-position taxonomy and the four principles] The central claim that the accountability boundary often must not move (while the interface boundary can) is load-bearing for the taxonomy and the four principles, yet the manuscript supplies no explicit operational criteria, decision rules, or measurement approach for classifying a given task or workflow into a regime where accountability must remain fixed versus one where it can migrate into prompts or agent instructions. This renders the distinction an assertion rather than a testable discriminator.

    Authors: We accept that the manuscript would be strengthened by greater specificity. The current version grounds the distinction in the cited theories and illustrates it with domain examples, but does not supply a classification rubric. In revision we will add a dedicated subsection that articulates three operational criteria: (1) whether the task requires professional or regulatory sign-off that cannot be transferred without altering liability allocation; (2) whether an immutable, auditable evidence trail is mandated by statute or professional standards; and (3) whether the workflow is tightly coupled to a trusted system of record whose integrity is cospecialized with the firm’s identity. These criteria will be presented as a decision checklist with concrete examples drawn from accounting, law, and healthcare. While the paper remains a conceptual contribution rather than an empirical measurement exercise, the added rubric will make the taxonomy more prescriptive and address the concern that the boundary distinction is asserted rather than operationalized. revision: yes

  2. Referee: [discussion applying Eisenmann, Parker, and Van Alstyne's platform envelopment framework] The argument that orchestrators operating through open protocols acquire envelopment power even as technical interoperability improves rests on the platform envelopment framework, but the manuscript does not address how the accountability boundary interacts with envelopment dynamics in concrete cases (e.g., whether regulated workflows in healthcare create a natural barrier or merely delay envelopment).

    Authors: We agree this interaction deserves explicit treatment. In the revised manuscript we will expand the envelopment section to show how the accountability boundary modulates envelopment speed. We will argue that in domains with strong accountability regimes (e.g., healthcare), regulated workflows and liability structures create a temporary barrier because they require cospecialized accountability assets that general-purpose orchestrators cannot replicate through prompts alone. We will contrast this with lower-accountability tasks where envelopment can proceed more rapidly. The revision will note that the barrier may only delay rather than prevent envelopment if rule debt accumulates, and will illustrate the point with the manuscript’s existing healthcare and procurement examples. This addition preserves the original theoretical claims while clarifying the dynamic role of accountability assets. revision: yes

Circularity Check

0 steps flagged

No circularity: derivation applies independent external frameworks

full rationale

The paper's argument distinguishes interface boundaries from accountability boundaries and derives a three-position taxonomy plus four principles by applying Coase's theory of the firm, Eisenmann et al.'s platform envelopment framework, and Teece's analysis of complementary assets. These are cited as established external references whose content does not depend on the present manuscript. No equations, fitted parameters, self-definitional constructs, or load-bearing self-citations appear in the provided text; rule debt is introduced as a new formalization rather than a renaming or reduction of prior inputs. The derivation therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 1 invented entities

The paper relies on three established economic theories as background assumptions and introduces one new conceptual entity (rule debt) without any fitted parameters or empirical data.

axioms (3)
  • domain assumption Coase's theory of the firm
    Invoked to analyze firm boundaries in the presence of AI agents reducing transaction costs.
  • domain assumption Eisenmann, Parker, and Van Alstyne's platform envelopment framework
    Used to explain how orchestrators gain power through open protocols.
  • domain assumption Teece's analysis of complementary assets and appropriability
    Applied to argue that value capture resides in accountability assets.
invented entities (1)
  • rule debt no independent evidence
    purpose: Captures the future governance, maintenance, and accountability burden that accrues when business rules migrate from governed systems into prompts and agent instructions.
    New construct introduced to highlight a customer cost of headless architectures.

pith-pipeline@v0.9.0 · 5827 in / 1385 out tokens · 41633 ms · 2026-05-20T10:55:55.473886+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

15 extracted references · 15 canonical work pages

  1. [1]

    Coase, R. H. 1937. The Nature of the Firm. Economica 4, 16, 386–405. https://www.jstor.org/stable/2626876

    work page arXiv 1937

  2. [2]

    Boudreau, K. 2010. Open Platform Strategies and Innovation: Granting Access vs. Devolving Control. Management Science 56, 10, 1849–1872. https://doi.org/10.1287/mnsc.1100.1215

    work page doi:10.1287/mnsc.1100.1215 2010

  3. [3]

    Eisenmann, T., Parker, G., and Van Alstyne, M. 2011. Platform Envelopment. Strategic Management Journal 32, 12, 1270–1285. https://doi.org/10.1002/smj.935

    work page doi:10.1002/smj.935 2011

  4. [4]

    Model Context Protocol. 2025. Specification. https://modelcontextprotocol.io/specification/2025-11- 25

    work page 2025

  5. [5]

    Teece, D. J. 1986. Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing and Public Policy. Research Policy 15, 6, 285–305. https://doi.org/10.1016/0048-7333(86)90027-2

    work page doi:10.1016/0048-7333(86)90027-2 1986

  6. [6]

    AS 1215: Audit Documentation

    Public Company Accounting Oversight Board. AS 1215: Audit Documentation. https://pcaobus.org/oversight/standards/auditing-standards/details/AS1215

    work page

  7. [7]

    AS 1105: Audit Evidence

    Public Company Accounting Oversight Board. AS 1105: Audit Evidence. https://pcaobus.org/oversight/standards/auditing-standards/details/AS1105

    work page

  8. [8]

    Cunningham, W. 1992. The WyCash Portfolio Management System. OOPSLA ’92 Experience Report. https://dl.acm.org/doi/10.1145/157709.157715

    work page doi:10.1145/157709.157715 1992

  9. [9]

    American Bar Association. 2023. Predictive Coding Technologies for the Modern Lawyer. Litigation News. https://www.americanbar.org/groups/litigation/resources/litigation- news/2023/summer/predictive-coding-technologies-modern-lawyer/

    work page 2023

  10. [10]

    Securities and Exchange Commission, Division of Investment Management

    U.S. Securities and Exchange Commission, Division of Investment Management. 2017. Robo- Advisers. IM Guidance Update No. 2017-02. https://www.sec.gov/investment/im-guidance-2017- 02.pdf

    work page 2017

  11. [11]

    Parker, G., Van Alstyne, M., and Jiang, X. 2017. Platform Ecosystems: How Developers Invert the Firm. MIS Quarterly 41, 1, 255–266. https://aisel.aisnet.org/misq/vol41/iss1/15/

    work page 2017

  12. [12]

    Z., and Iqbal, R

    Telang, R., Hydari, M. Z., and Iqbal, R. 2026. To Scale AI Agents Successfully, Think of Them Like Team Members. Harvard Business Review (March 2026). https://hbr.org/2026/03/to-scale-ai- agents-successfully-think-of-them-like-team-members Appendix A. Bilateral Contracting in Vertical AI: What Changes When There Is a Partnership? The main article focuses ...

    work page 2026

  13. [13]

    Williamson, O. E. 1985. The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. Free Press

    work page 1985

  14. [14]

    Grossman, S. J. and Hart, O. D. 1986. The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration. Journal of Political Economy 94, 4, 691–719

    work page 1986

  15. [15]

    and Moore, J

    Hart, O. and Moore, J. 1990. Property Rights and the Nature of the Firm. Journal of Political Economy 98, 6, 1119–1158

    work page 1990